IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 19, NO.

2, APRIL 2004 595

Method for Calculating the Ground Resistance

of Grounding Grids Using FEM
J. A. Güemes, Member, IEEE, and F. E. Hernando

Abstract—Taking as given that ground resistance is indepen- people and equipment. The safety, reliability, and correct oper-
dent of earth fault current, this paper presents a new way of calcu- ation of electrical power systems depend on the quality of the
lating ground resistance of grounding grids using the finite-element design and construction of their grounding grids.
method (FEM). Ground resistance is calculated by means of three
consecutive tests, with an acceptable number of elements, and fast The main objectives of a grounding grid are:
resolution. The results of the proposed method are compared with • to protect personnel against electrical risks by limiting
those measured experimentally or determined by other authors. the overvoltages to which they may be exposed if ground
Once ground resistance and earth fault current are known, grid
faults occur in substations or generating stations;
potential and touch-and-step voltages are calculated. The main ad-
vantage of the method proposed here is that it permits us to analyze • to ensure safety and continuity of electrical equipment by
symmetrical and nonsymmetrical grids of any shape in uniform, limiting the overvoltages that can appear under extreme
two-layer, and multilayer soils. The method developed in this paper operation conditions or in case of an accident;
constitutes a fast-resolution industrial application, with acceptable • to ensure correct operation of equipment and electrical
results, for calculating the ground resistance of grounding grids of
protection devices by enabling ground faults to be detected
any shape.
and actions selected to disconnect those zones of the elec-
Index Terms—Earth potentials, grounding grids, ground resis- trical installation where faults occur.
tance, modeling, simulation.
This paper presents a method of considerable help to
electrical power engineers, as it enables them to calculate the
I. INTRODUCTION ground resistance of grounding grids of any shape (square,
rectangular, triangular, T-shaped, L-shaped, etc.), in uni-
I NCREASING social sensitivity with regard to electrical
risks for people and more stringent technical demands in
electrical protection systems make it necessary to carry out an
form, two-layer, and multilayer soils and soils with finite
heterogeneities or horizontally nonstratified multilayer soil
exact study of grounding grids. structures, via three consecutive quickly performed tests. The
Grounding grids are basically made up of round steel rods method proposed, which has been tried and tested in practice,
forming two-dimensional (2-D) grids (normally square or rect- is independent of the grounding grid shape, the soil structure,
angular in shape), buried near the soil surface (0.4–0.6 m). In the grid potential, and the boundary condition defined in the
grids made up of meshes and ground rods, the ground rods are simulation of grounding grid behavior.
only effective if a significant portion of their length is in contact
with a low resistivity soil [1]. II. BACKGROUND
Ground resistance in electrical power systems is defined as
, where is the electric potential of the grounding grid, and One important step in determining the size and basic layout
the earth fault current (electric current that flows into the earth of a grounding system for an ac substation is to estimate the
through the grounding grid). ground resistance of the grounding grid.
The ground resistance for a given fault current determines Calculation methods for determining the ground resistance of
what hazardous voltages exist inside or around the substation grounding grids consider the following simplifying hypotheses.
or generating station. Consequently, the ground resistance is an • The soil is an infinite medium which is flat, isotropic, and
important technical parameter which is related to the safety of stratified in layers of uniform thickness.
people and equipment. If the ground resistance is too high and • The laws of electromagnetism may be applied for calcu-
an earth fault occurs, personnel may be killed or injured and lating the ground resistance and potential distribution in
equipment may be damaged. the soil.
Grounding grids are, without doubt, the most important part • Grounding grid rods are assumed to be linear, intercon-
of an electrical system from the point of view of the safety of nected, and buried close to the soil surface.
• Grounding grid behavior at rated working frequency can
be determined using electromagnetic field analysis tech-
Manuscript received August 26, 2003.
J. A. Güemes is with the Electrical Engineering Department, University of niques for stationary fields (propagation time is rejected).
the Basque Country, Bilbao 48012, Spain (e-mail: [email protected]). The different calculating methods are based on the determi-
F. E. Hernando is with the Electrical Engineering Department, University of nation of grid voltage or “grid capacitance.”
the Basque Country, Bilbao, Spain, and also with the Instalaciones Eléctricas
Scorpio, Loiu 48180, Spain (e-mail: [email protected]). The first method is based on determining the grounding grid
Digital Object Identifier 10.1109/TPWRD.2004.824761 potential, usually by means of the image method [2], [3].
0885-8977/04$20.00 © 2004 IEEE
596 IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 19, NO. 2, APRIL 2004

The second method is based on determining the electrode

capacitance on the basis of the relation between electric charge
and potential, once the electric field in the soil has been
calculated.
Once we know the grounding grid potential and the earth fault
current, the ground resistance of the grounding grid is calculated
by applying Ohm’s law. Fig. 1. Ground resistance.
Simple formulas for calculating ground resistance in square
grounding grids have been proposed by Dwight et al. [4]–[7], in
Moreover, voltage measurements in the soil prove that the
rectangular grids by Schwarz [8], and in grounding grids of any
potential value at each point of the soil depends on its position
shape by Thapar et al. [9].
with respect to the grounding grid, where the law of variation of
Current grounding grid calculation methods determine
ground resistance–in terms of length–and starting from a certain
ground resistance and the step and touch voltages using
distance from the center of the grid, is an increasing exponential
different mathematical techniques, applying hypotheses that
function (see Fig. 2).
allow us to model the real system in theoretical systems with
Taking as given that potential distribution is exponential, the
comparable results. These studies are performed generally
first stretch of resistance is not considered (see Fig. 2), the
for symmetrical grounding grids with uniform, two-layer, or
following is fulfilled:
multilayer soils [10]–[13].
Recently, studies based on the finite-element method (FEM)
have been used to calculate ground resistance of grounding
grids. The first simulation studies of grounding grid behavior
using the FEM were based on calculating ground resistance for
an arbitrary grid potential (once the grid current is known). The
grid current for the grid potential set is determined by means of (1)
a current flow analysis. Once the current is calculated, ground
resistance is determined as the quotient between the voltage set From quotient , the following expression is obtained:
and the current calculated [14].
Model size selection was difficult in this method (earth dis- (2)
tance to be considered starting from the grounding grid), and
this conditioned the value of the calculated ground resistance.
Similarly
To decrease the error of the ground resistance calculated, elec-
trical power engineers were forced to analyze models of dif-
ferent sizes with a high number of nodes. Due to the low levels of
accuracy of the results and the long calculation times required,
the method is not very feasible. Another additional problem is (3)
the boundary conditions to be applied for the size of the model
considered. By means of three consecutive tests, for earth distances ,
As a result of the difficulties of the method outlined above, , and considered from the center
a new method has been developed that enables ground resis- of the grid, the resistances , , and are determined.
tance to be determined starting from the dissipated power or From the resistance values thus calculated, the value of the
from the stored energy (by the electric field) in the model. This constant is determined
method has the additional advantage of being independent of
the boundary condition, shape, and size of the grid and soil (4)
structure.
The method presented in this paper may prove highly useful where can be obtained by subtracting of the resistance cal-
in determining precise formulas for calculating ground resis- culated in the third test from that in the second and by sub-
tance in different kinds of grounding grid, with no need to build tracting of the resistances calculated in the second test from that
and measure large numbers of grounding grids or study scale in the first.
models [15]. Finally, the ground resistance is determined by means of the
equation
III. METHOD
Admitting that the ground resistance depends solely on the
shape and size of the grounding grid, the soil structure (conduc- (5)
tivity and number of layers) and the soil length considered, and
is independent of grid potential (earth fault current), ground re- Since the distribution of the resistance, as a function of dis-
sistance can be obtained as the sum of a set of resistances in tance, is an increasing exponential function, parameter “ ” (3)
series (see Fig. 1). is negative and consequently is positive and smaller than one.
GÜEMES AND HERNANDO : METHOD FOR CALCULATING GROUND RESISTANCE 597

where
surface that completely surrounds an electrode;
any trajectory that joins the two electrodes;
electrical conductivity;
electric field.
From (10), it is deduced that resistance depends only on ge-
ometry and electrical conductivity.
In general, it is more advisable to calculate resistance (5)
starting from dissipated power, by means of the following
expression:
Voltage
(11)
Dissipated power
At the same time, the dissipated power is determined by

(12)

Fig. 2. Voltage. B. Electrostatic Analysis

For an arbitrary geometry, the capacity of a dielectric medium
Accordingly, it is deduced that the series can be expressed in terms of electric field by means of the fol-
is convergent lowing expression:
If
(13)
(6)

and if both sides of the above equation are multiplied by (6), where is the electric permittivity.
the following expression is obtained: Moreover, the stored energy by the electric field is given by
(7)
(14)
Subtracting (7) from (6) and calculating the limit when n
tends to infinity, the following expression is obtained: For the same configuration with a homogenous ohmic con-
ductor or for a linear dielectric, the product –in agreement
(8) with (10) and (13)–is a constant independent of the geometry

Finally, substituting this expression in (5), we have (15)

Finally, considering (14) and (15), the following expression

(9)
is obtained:
This equation allows us to calculate the ground resistance of a
(16)
grounding grid once the resistances , , and are known.
Taking as given that ground resistance is independent of earth This equation allows us to determine the ground resistance
fault current, the tests can be carried out for any grid potential. once the stored energy by the electric field is known.
The resistance in each stretch as indicated above is calculated
by using the FEM and performing a current flow analysis or an
IV. MODEL
electrostatic analysis.
Once the ground resistance had been calculated, the FEM can The model comprises a volume with a rectangular cross sec-
be analyzed again using the actual grid potential (calculated as tion and depth , and an air layer above the soil of height .
the product of the actual fault current and the ground resistance In that volume, the grounding grid is located at a depth . The
calculated), to obtain the potential distribution in the soil and ground must be divided into layers according to the electrical
the step and touch voltages. characteristics of the soil. Before defining the model, it is very
important to check whether the grounding grid presents elec-
A. Current Flow Analysis trical symmetries.
For an arbitrary geometry, the resistance between two elec- The finite-element mesh is made up of 8-node elements. It is
trodes can be expressed in terms of electric field by [16] important that the elements are of uniform size in the proximity
of the rods of the grounding grid.
Properties are defined by means of the electric conductivity
(10)
of the air and of each soil layer (current flow analysis) or by
598 IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 19, NO. 2, APRIL 2004

TABLE I
GROUND RESISTANCE

TABLE II
GROUNDING GRIDS. Resistivity of the soil = 100 1 m,
Diameter of conductor := 10 mm, Depth of burial = 0:5 m,
Number of meshes = n, Dimensions of grid = a; b.

Fig. 3. Grounding grid.

TABLE III
GROUND RESISTANCE VALUES

Fig. 4. Finite-element mesh.

TABLE IV
MISTAKES (IN PERCENT)
means of the electric permittivity of the air and of each soil layer
(electrostatic analysis).

V. APPLICATION EXAMPLES
In this section, the method described above is applied to cal-
culate the ground resistance of a 12 8 m, 2-mesh, 6-ground
rod grounding grid buried at a depth of 0.6 m in two-layer soil.
The first layer has a thickness of 0.2 m and a conductivity of
m and the second layer has a conductivity
of 1/22 m .
Fig. 3 shows a plan view of the grounding grid indicated.
Given the electrical symmetries, it suffices to study half of the
model or , where is the soil length
considered in each simulation of grounding grid behavior.
Fig. 4 shows the finite-element mesh used.
Table I shows the ground resistance obtained once the finite-
element models have been analyzed.
is given by

(17)

From (9), ground resistance is 0.988 . The actual ground Fig. 5. Ground resistance variation.
resistance measured for this grounding grid was 1 .
Fig. 5 shows the variation in ground resistance with the soil Fig. 6 shows the potential distribution on the soil surface and
length considered. also in a plane with a depth of 0.6 m (grounding grid plane).
GÜEMES AND HERNANDO : METHOD FOR CALCULATING GROUND RESISTANCE 599

Fig. 6. Distribution of potentials (Earth fault current = 1200 A).

The method presented in this paper has also been applied for The method takes the arbitrary grid potential as its initial data
calculating the ground resistance of the grounding grids shown and determines the ground resistance of the grounding grid by
in Table II. first calculating the electric field and dissipated power.
Table III shows the ground resistance values obtained with The results of the method presented in this paper are good in
the method explained in this paper and those calculated in [9] comparison with those obtained experimentally or calculated by
using the methods developed by Thapar et al. other authors.
Taking into consideration these last results, we can see that The method presented in this paper may be highly useful in
the difference between the values calculated and those proposed determining precise formulae applied to calculating ground re-
in [10] is no more than 10% (see Table IV). sistance in different kinds of grounding grid.
Once the ground resistance and the earth fault current are
known, the potentials in the nodes of the soil surface and the
VI. CONCLUSION touch-and-step voltages can be calculated.
A method for calculating the ground resistance of grounding
grids using the FEM has been presented in this paper. REFERENCES
The method presented is based on the independence of [1] L. Hyung-Soo, K. Jung-Hoon, F. P. Dawalibi, and M. Jinxi, “Efficient
ground resistance with respect to earth fault current. ground grids designs in layered solids,” IEEE Trans. Power Delivery,
vol. 13, pp. 745–751, July 1998.
The method proposed, which has been tried and tested in [2] P. J. Lagace, J. L. Houle, Y. Gervais, and D. Mukhedkar, “Computer
practice, is independent of the shape of the grounding grid, the aided design of a toroidal ground electrodes in a two layer soil,” IEEE
soil structure (uniform, two-layer, and multilayer soils), the grid Trans. Power Delivery, vol. PWRD-2, pp. 744–749, July 1987.
potential, and the value of the boundary conditions defined in [3] P. J. Lagace, D. Mukhedkar, H. H. Hoang, and H. Greiss, “Evaluation
of the effect of vertical faults on the voltage distribution around HVDC
the finite-element model. This is a useful method for engineers electrodes using a super-computer,” IEEE Trans. Power Delivery, vol.
engaged in the design and analysis of grounding systems. 5, pp. 1309–1313, July 1990.
600 IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 19, NO. 2, APRIL 2004

[4] H. B. Dwight, “Calculation of resistance to ground,” Elect. Eng., vol. [16] M. Zahn, “Teoría electromagnética,” in Interamericana: Nueva editorial
55, pp. 1319–1328, Dec. 1936. interamericana, S.A. de C.V., México D.F., 1987, pp. 169–177.
[5] P. Laurent, “General fundamentals of electrical grounding techniques,”
Le bulletin de la Société Francaise des Electiciens, vol. 1, pp. 368–402,
July 1951.
[6] J. Niemann, “Unstellung von Hochstspannungs-Erdungsalagen Aufden
betrieb mit starr geerdetem sternpunkt,” Electrotech. Z., vol. 73, pp.
J. A. Güemes-Alonso (M’99) was born in La Rioja,
333–337, May 1952.
Spain, in 1952. He received the industrial technical
[7] J. G. Sverak, “Simplified analysis of electrical gradients above a ground
engineering and industrial engineering degrees
grid: Part I – How good is the present IEEE method?,” IEEE Trans.
in electrical engineering, and the Ph.D. degree in
Power App. Syst., vol. PAS-103, pp. 7–25, Jan. 1984.
industrial engineering from the University of the
[8] S. J. Schwarz, “Analytical expression for resistance of grounding sys-
Basque Country, Bilbao, Spain, in 1975, 1979, and
tems,” AIEE Trans., pt. III-B, vol. 73, pp. 1011–1016, 1954.
1985, respectively.
[9] B. Thapar, V. Gerez, A. Balakrishnan, and D. A. Blank, “Evaluation of
Currently, he is Catedrático de Escuela Universi-
a grounding grid of any shape,” IEEE Trans. Power Delivery, vol. 6, pp.
taria with the Electrical Engineering Department of
640–647, Mar./Apr. 1991.
the Basque Country University. His research interest
[10] F. Dawalibi and D. Mukhedkar, “Optimum design of substation
is the design of devices and electrical machines.
grounding in two-layer earth structure- Part. I, Analytical study,” IEEE
Trans. Power App. Syst., vol. PAS-94, pp. 252–261, Mar./Apr. 1975.
[11] F. P. Dawalibi and D. Mukhedkar, “Optimum design of substation
grounding in two-layer earth structure–Part II, Comparison between
theoretical and experimental results,” IEEE Trans. Power App. Syst.,
vol. PAS-94, pp. 262–266, Mar./Apr. 1975. F. E. Hernando-Fernández was born in the
[12] H. R. Seedher, J. K. Arora, and B. Thapar, “Finite expressions for com- Province of Segovia, Spain, in 1955. He received the
putation of potential in two layer soil,” IEEE Trans. Power Delivery, vol. industrial engineering degree in electricity from the
PWRD-2, pp. 1098–1102, Oct. 1987. Escuela Técnica Superior de Ingenieros (E.T.S.I.I.),
[13] H. S. Lee, J. H. Kim, F. P. Dawalibi, and J. Ma, “Efficient ground grid University of the Basque Country, Bilbao, Spain, in
designs in layered soils,” IEEE Trans. Power Delivery, vol. 13, pp. 1980.
745–751, July 1998. Currently, he combines his work as Profesor Tit-
[14] J. A. Güemes and F. E. Hernando, “Cálculo de redes de tierra utilizando ular de Escuela Universitaria with the Electrical En-
el método de elementos finitos. Comparación de resultados,” in Proc. gineering Department at the Escuela Universitaria de
IEEE Andrean Región Int. Conf., vol. 2, 1997, pp. 902–906. Ingenería Técnica industrial of Bilbao (University of
[15] B. Thapar and S. L. Goyal, “Scale model studies of grounding grids the Basque Country), Spain, and as an Industrial En-
in nonuniform soils,” IEEE Trans. Power Delivery, vol. PWRD-2, pp. gineer with “Instalaciones Eléctricas Scorpio, S.A.,” Bilbao, Spain, a company
1060–1066, Oct. 1987. which designs and builds low- and high-voltage electrical installations.